Endpoint pricing
What every SocialCrawl endpoint costs in credits, searchable by platform, path, or tier
Every request is billed in credits. Most endpoints cost 1 credit per successful call. A minority cost more because the upstream is slower, returns more data, or runs extra compute, and a set of bundle endpoints price per request instead of per call.
Costs here come from the same source the API charges against, so a number on this page is the number on your invoice.
The tier ladder
| Tier | Credits | Endpoints | Typical use |
|---|---|---|---|
| Standard | 1 | 272 | Profiles, posts, comments, basic search |
| Advanced | 5 | 172 | Trending feeds, ad libraries, audience analytics |
| Premium | 10 | 21 | Video transcripts, LinkedIn people and job search |
| Custom | varies | 90 | Bundle and fan-out endpoints |
555 active endpoints across 64 platforms. Endpoints that are temporarily unavailable are not listed: those return 503 and cost nothing.
Custom-priced endpoints
Custom-priced endpoints do not sit on the 1/5/10 ladder. They price per request, because one call fans out to a variable number of upstream calls: search pages, comment threads, brands, or extra platforms.
Those endpoints show a range such as 2-200 (metered). You are always charged inside that range and never below its floor. Filter to Custom below to see all 90 of them with their ranges.
Find an endpoint
555 of 555 endpoints
How a charge is applied
| Case | What happens |
|---|---|
| Successful call | The listed credits are deducted when the response is dispatched. |
Cache hit ("cached": true) | 0 credits. See Credits. |
| Upstream error, 5xx, empty-body 404 | The deduction is refunded automatically. See Credits. |
| Idempotent retry | Charged once per Idempotency-Key for 24 hours. A replay reports credits_used: 0 and idempotent_replay: true. |
Balance check (GET /v1/credits/balance) | 0 credits, always. |
Two cases work differently and are worth reading before you leave one running.
Async jobs and browser sessions hold, then refund. POST /v1/web/crawl, POST /v1/web/batch-scrape, POST /v1/web/agent and POST /v1/web/sessions deduct an upper-bound hold before the upstream call, then refund the unused portion when the job settles or the session closes. Your balance drops at submit time by the hold shown in the table above, not by a flat per-call price. A job that fails upstream is refunded in full.
Monitors bill on every check. Creating a monitor is free. Each scheduled check bills its upstream cost plus 1 credit for orchestration, and keeps billing on your cadence until you pause or delete it. At the tightest 5-minute cadence that is 288 checks a day. The create response returns credits_booked, the upstream cost of one check, so budget (credits_booked + 1) x checks per month.
On billed responses, credits_remaining and X-Credits-Remaining are authoritative. Free cache hits return credits_remaining: null and omit the header because they do not establish an authoritative post-request balance. Idempotent replays return the current balance when their lookup succeeds; otherwise they return credits_remaining: null and omit the header. Routes that explicitly model an unavailable balance, including endpoint-parameter validation and best-effort metadata, omit both fields when their balance read fails.
Related
Credits
How charging, refunds, cache hits, and balance checks work.
Which endpoint should I use?
The endpoint pairs developers confuse, compared side by side with prices.
Rate limits
Per-key request and concurrency limits, and what a 429 looks like.
API reference
Every endpoint with parameters, response schemas, and a playground.